High hematocrit and the donation reflex

What testosterone actually does to red cells Strong evidence

Almost everyone on testosterone sees hematocrit rise. The usual explanation stops at "testosterone stimulates red blood cell production". True, but not useful, because it does not explain the thing people find confusing: why the number climbs for months and then sits there instead of coming back down.

The clearest account comes from a trial that measured the machinery rather than the outcome. Bachman and colleagues gave older men testosterone gel or placebo for six months and tracked erythropoietin, hepcidin and ferritin alongside the blood count. Hemoglobin and hematocrit rose by roughly 7 to 10%. Erythropoietin went up. Hepcidin and ferritin went down, at one month and at three months. Hepcidin is the master regulator of iron availability, so suppressing it releases iron into circulation. The marrow is handed both the signal and the raw material at the same time.

Then the interesting part. By six months, erythropoietin and hepcidin had drifted back toward baseline while hemoglobin and hematocrit stayed elevated. Under normal feedback, a high red cell mass should suppress erythropoietin below baseline. It did not. The authors described this as a new erythropoietin set point: testosterone has not applied a temporary push, it has renegotiated what your body considers normal and then stopped arguing.

That single finding explains most of the practical confusion. It is why hematocrit does not correct itself while the dose is unchanged. It is why a first blood draw at three months tends to underestimate where you will actually land. And it is why the meaningful intervention is the one that changes the set point, which is the dose, rather than the one that empties the bucket and leaves the setting alone.

Where 54% came from and what it is not Strong evidence

The Endocrine Society's 2018 clinical practice guideline is where the numbers everyone quotes come from. It recommends checking hematocrit at baseline, again at three to six months and annually after that. It advises against starting testosterone in a man whose baseline hematocrit is already above 48%, or above 50% at high altitude, calling that a relative contraindication. And above 54% it recommends stopping therapy until the value normalizes, then restarting at a reduced dose, while investigating causes of hypoxia such as sleep apnea.

That is a sound monitoring schedule and it is worth following. But it is worth being precise about what 54% is. It is a threshold at which a guideline committee judged the balance of caution tips toward acting. It is not a measured boundary where thrombotic risk switches on. The guideline says so itself: the hematocrit level at which the risk of neuro-occlusive or cardiovascular events increases is not known. Nobody randomized men to 53% versus 55% and counted clots.

Two practical consequences follow. A reading of 52% is not a license to ignore a rising trend. A single reading of 55% is not an emergency. And because hematocrit is a ratio, the fraction of blood volume made up of red cells, it moves when the plasma volume moves. Arriving at a blood draw dehydrated, after a fasted morning, a hot session or a heavy night, inflates it. That is the single most common reason a frightening number turns out to be a hydration artifact. Before acting on a first high result, repeat it properly hydrated. The bloodwork guide covers draw-day timing for the rest of the panel too.

Does thick blood actually hurt you? Mixed

This is where the honest answer is more interesting than the confident one. The intuitive model is that thicker blood means more viscosity, more strain and more arterial events, which is the framing most articles on this subject use. The largest trial to look at testosterone and cardiovascular outcomes does not support that model. It does support a different one.

TRAVERSE randomized 5,246 middle-aged and older men with symptomatic hypogonadism and either established cardiovascular disease or a high burden of risk factors to testosterone gel or placebo. For major adverse cardiovascular events, the composite of cardiovascular death, non-fatal myocardial infarction and non-fatal stroke, testosterone was non-inferior to placebo. The classic arterial fear was not what the trial found.

What it did find, in the secondary safety data, was more pulmonary embolism, more atrial fibrillation, more non-fatal arrhythmia and more acute kidney injury in the testosterone arm. Those are real signals and they point at venous thromboembolism and rhythm rather than at plaque. So the risk is not imaginary. It is also not the risk people usually name. If you are going to worry about something, a clot in a vein is better supported by the evidence than a heart attack.

It is worth noting what TRAVERSE was not designed to answer. It did not stratify outcomes by hematocrit, so it cannot tell you that a man at 55% is at more risk than a man at 48%. The link between the number on your panel and the events in the trial remains inferred rather than demonstrated.

Outside the trials the association is real even if causality is not settled. The same 2024 review that is sceptical about phlebotomy is clear about this. Its own summary opens by saying evidence shows an increased thrombotic risk associated with testosterone-induced erythrocytosis. The strongest piece is a propensity-matched cohort drawn from a database of around 74 million records, in which hypogonadal men whose hematocrit reached 52% or more on testosterone had more cardiac events plus more venous clots than those whose did not, at an odds ratio of 1.35. At a 50% cutoff there was no significant difference. The review also notes why the trials cannot see this: they exclude men with a high baseline hematocrit then dose-adjust anyone who becomes erythrocytotic, so the men presumably at highest risk are the ones missing from the data.

Why donating blood is not the automatic fix Mixed

Ask any forum what to do about high hematocrit and you will be told to donate blood. It is the most confidently repeated advice in the space and it rests on the thinnest evidence of anything on this page.

A 2024 review in Endocrine Connections by Bond, Verdegaal and Smit examined exactly this question and concluded that evidence supporting the efficacy or safety of therapeutic phlebotomy in testosterone-induced erythrocytosis is lacking. The guideline endorsements exist, but they are extrapolated from polycythemia vera, where phlebotomy has a genuine evidence base. Polycythemia vera is a clonal marrow disease driven by a JAK2 mutation with suppressed erythropoietin and a documented thrombotic risk. Testosterone-induced erythrocytosis is a secondary, hormone-driven rise with erythropoietin that is normal or high. Same number on the panel, different disease underneath. Evidence does not transfer for free between them.

The review goes further and argues phlebotomy may not be neutral. Removing blood repeatedly lowers tissue oxygen tension and eventually depletes iron stores. Low iron reduces prolyl hydroxylase activity, which is what marks hypoxia-inducible factor for degradation. So HIF accumulates, on a pathway that has prothrombotic effects of its own, entirely independent of hematocrit. The plausible outcome of an aggressive donation schedule is a man whose hematocrit is a little lower, whose iron stores are empty and whose actual risk has not moved in the direction he thinks.

There is a practical version of this that shows up in bloodwork constantly. Serial donation without monitoring ferritin produces iron-deficient erythrocytosis: hematocrit still high because the set point has not changed, ferritin on the floor and a man who feels considerably worse than his hematocrit suggests he should. He then blames the testosterone.

None of this makes donation wrong. It makes it a considered decision rather than a reflex, taken with ferritin in view, after the interventions with a mechanism behind them have been tried. The review's own recommendation is to prioritize dose reduction and to address contributors like smoking and sleep apnea first and to treat phlebotomy as shared decision-making that acknowledges the uncertainty.

What actually lowers it Mixed

In rough order of how well the mechanism holds up:

The frequency question is the one people most want a hard answer to and it does not have one. If you are going to test it on yourself, change one variable at a time and allow three months, because red cell turnover is slow and a single draw six weeks after a change is noise rather than a result. That is an experiment for a hematocrit that is drifting, not one already over 54%. Above that line the guideline's answer is to stop and reassess with a clinician rather than to wait a quarter. The injection schedule guide covers how frequency changes the curve. TRT protocols covers the dose side.

Read the panel, not the number Strong evidence

Three values tell the story together. Looking at one of them alone is how people reach the wrong conclusion.

Hematocrit is a ratio and therefore hydration-sensitive. Hemoglobin measures the oxygen-carrying protein itself and moves less with plasma volume, so a hematocrit that looks alarming next to an unremarkable hemoglobin usually means you arrived dry. Ferritin is the one most people never order and the one that changes the interpretation most: it tells you whether iron stores are intact, whether hepcidin suppression has liberated iron as expected and above all whether a donation habit has quietly emptied you out.

Timing matters as much as the panel. A trough draw before the next injection and a peak draw a couple of days after are different numbers. Comparing one to the other across visits produces a trend that is not real. Draw at the same point in the cycle every time. That principle runs through the whole monitoring panel. It is the difference between tracking a protocol and collecting anecdotes.

Cycle doses are a different question Emerging

Everything quantified above comes from replacement-dose trials in hypogonadal men. It is worth stating plainly that applying it to a gram a week of multiple compounds is extrapolation.

A 2024 systematic review and evidence-gap mapping covering 22 studies of 1,023 bodybuilders, 662 of them anabolic steroid users, published between 1987 and 2022, found that the hematological parameters rest on only one or two studies. That is a gap, not a finding. The direction is not in doubt: supraphysiological androgen doses push hematocrit higher and further than replacement doses. What does not exist is a calibrated dose-response curve, or a threshold validated in that population, or an outcome trial. Anyone quoting you a precise figure for what 750 mg a week does to your hematocrit is quoting a forum.

Practically, that argues for measuring more often rather than reasoning more confidently. If you are running compounds well above replacement, the interval between panels should be shorter than the guideline's annual cadence. Hematocrit should be read alongside blood pressure and lipids rather than on its own. The cycle calculator flags which compounds in a stack carry a hematocrit signal. The compound profiles cover the structural reasons why. Neither can tell you what your own number will do, which is the entire argument for measuring it.